journal_of_animal_geneticsfandomcom-20200213-history
The myostatin gene: an overview of mechanisms of action and its relevance to livestock animals
D. Aiello K. Patel E. Lasagna First published: 20 August 2018 https://doi.org/10.1111/age.12696 Introduction Myostatin Myostatin (MSTN), also known as growth and differentiation factor 8 (GDF8), is one of the major regulators of skeletal muscle development (Beyer et al. 2013). The MSTN gene is highly conserved among mammalian species, and it acts in an almost unique manner to reduce muscle size. MSTN‐deficient animals display an increase in skeletal muscle mass known as double muscling (DBM). Mutations in MSTN have been described in numerous species including dog (Mosher et al. 2007), sheep (Kijas et al. 2007), cattle (Grobet et al. 1997), pig (Stinckens et al. 2008) as well as in one human (Schuelke et al. 2004). Myostatin signalling pathway and its control of skeletal muscle development Myostatin is expressed in many tissues (including the mammary gland) but most prominently in skeletal muscle (Ji et al. 1998). The MSTN gene has been highly conserved throughout evolution and comprises three exons and two introns. In all species reported in this review, MSTN exons code for a 375‐amino‐acid (aa) latent protein that undergoes significant post‐translational modification in order to become biologically active (Wolfman et al. 2003). Firstly, the polypeptide undergoes intracellular homodimerization through the formation of disulphide bonds. Thereafter, it is cleaved to form the N‐terminal propeptide region and the C‐terminal mature region. The 12‐kDa C‐terminal mature fragment of MSTN initiates an intracellular signalling cascade through its ability to bind and activate the activin type II receptor at the cell surface (ActRIIB and to a lesser extent ActRIIA). Subsequent autophosphorylation of ActRIIB leads to the recruitment and activation of the low affinity type I receptor for Activin ALK‐4 or ALK‐5. Activated type I receptor kinase phosphorylates the transcription factors Smad2 and Smad3, allowing them to interact with Smad4 (co‐Smad) and to translocate to the nucleus to activate target gene transcription. Importantly, the activation of the MSTN receptor also inhibits Akt (protein kinase B) activity, a major determinant in muscle protein synthesis and cell proliferation. Enlargement of muscle fibre size, a process called fibre hypertrophy (or simply hypertrophy) is in large part controlled by Akt activity (Trendelenburg et al. 2009). Myogenic differentiation is a highly orchestrated sequential program that ultimately generates mature skeletal muscle. Highly proliferative muscle precursors that arise during embryogenesis differentiate into myoblasts. The commitment of the myogenic lineage is regulated by muscle regulatory factors (MRFs), a collective group of helix‐loop‐helix transcription factors, namely, MyoD, Myf5, Myogenin and MRF4 (Fig. 1). Additionally, exit from the cell cycle is a vital step during myoblast differentiation (Bryson‐Richardson & Currie 2008). Myostatin regulates muscle development at key points during the process of pre‐natal muscle development: muscle precursor proliferation, myoblast proliferation and differentiation. A study by Amthor et al. (2002) has shown that ectopic expression (in limb muscle) of MSTN down‐regulates Pax3, a key marker of proliferating muscle precursors (Amthor et al. 2002). Additionally, MSTN upregulates p21 expression, which ultimately inhibits proliferation of MyoD‐expressing myoblasts (Thomas et al. 2000). Of relevance to this review is the relationship between MyoD activity and the expression of MSTN. MyoD is an important regulator of MSTN expression during myogenesis. This is demonstrated by a critical role of E‐box motifs that have been identified in the MSTN promoter region; these motifs are known to be the binding sites for basic helix‐loop‐helix transcription factors (MRFs) (Hu et al. 2013). The interrelationship between MyoD and MSTN ensures that promiscuous differentiation mediated by an over‐active MyoD‐induced cascade is checked by the up‐regulation of MSTN. Therefore MSTN serves to limit the size of both the myoblast precursor (Pax3+/MyoD+) and myoblast (Pax3−/MyoD+) pools. Down‐regulating the expression of MSTN would lead to an expansion of both populations (Amthor et al. 1999). Examination of mouse development shows that muscle mass is determined by the ability of myoblasts to form fibres, a process that occurs in two phases: primary and secondary fibre formation. Matsakas et al. (2010) have shown an increase in the myoblast pool just before the fibre formation process in myostatin‐null mouse (Mstn−/−) embryos, which supports the development of extranumerary primary and secondary myofibres. Any programme that promotes an increase in fibre formation is called fibre hyperplasia or simply hyperplasia (Amthor et al. 2002). Therefore, the Mstn−/− mouse displays hyperplasia as a consequence of developing an increased number of mono‐nucleated muscle cells (Matsakas et al. 2010). Shortly before birth, muscle in Mstn−/− mice not only contain extra muscle fibres but also each fibre has undergone a small, albeit significant, increase in size (18%). However, this is not enough to explain why the muscles in these individuals often weigh two to three times more than their normal counterpart (Omairi et al. 2016). The resolution to this issue comes by examining the size of each muscle fibre in adult mice. This reveals that, in the mouse, the increased muscle mass has arisen due to a combination of a pre‐natal increase in the number of fibres (hyperplasia) and a precocious post‐natal increase (43%) in the size of each fibre (hypertrophy) (McPherron & Lee 1997). These studies are extremely insightful when attempting to determine the cellular mechanism underpinning DBM in large mammalian species harbouring a MSTN mutation (Elashry et al. 2012). They predict that for an animal to develop fibre hyperplasia and a small degree of hypertrophy as a consequence of a MSTN mutation, the gene must be normally expressed and properly translated into a mature form during pre‐natal development. However, in order to display significant fibre hypertrophy these conditions need to be satisfied during post‐natal life. If the mouse is taken as a guide, then changes in fibre number and small changes in fibre diameter (less than 20%) can be explained by pre‐natal action of MSTN. In cattle, very low levels of MSTN are detected from day 15 to day 29 embryos, and increased expression is detected from day 31 onwards (Kambadur et al. 1997). The increase of MSTN expression in bovine embryos is thought to occur at a gestational stage when primary myoblasts are starting to fuse and differentiate into myofibres. Therefore the null mutation in the bovine MSTN leads to hyperplasia. Double muscling phenotypes The term hypertrophy has often been used to describe the enlargement of muscle, at the gross anatomical level, displayed by members of large mammalian species. Mechanistically this term has been used loosely, as in many cases enlargement of muscle occurs solely through pre‐natal muscle hyperplasia without any post‐natal fibre hypertrophy. Double muscling in large animals has been reported in several species. Generally, muscle with a large superficial area tends to be the most enlarged, whereas deeper muscles tend to be reduced in size relative to normal muscle (Ouhayon & Beaumont 1968). Large commercially important DBM animals, especially cattle, have excellent conformation and an extremely high carcass yield, coinciding with a reduced internal organ mass (Fiems 2012). However, these animals are more susceptible to respiratory disease, urolithiasis, lameness, nutritional stress, heat and dystocia, resulting in lower robustness (Holmes et al. 1973). Also, reproductive performance can be influenced by hypertrophy: for example, in the South Devon breed, the gestation period for DBM calves is longer, resulting in offspring with higher birth weights than for normal calves, also evidenced by higher instances of dystocia with high mortality rates if births are unassisted; the findings highlighted therefore that the segregating alleles at the MSTN locus have significant effects on calving ease in this breed (Wiener et al. 2002). Double‐muscled cattle have shown signs of fatiguing faster than normal cattle during forced exercise, relating to metabolic acidosis because of a reduced blood circulation leading to a deficiency in the transport of oxygen and a reduction of aerobic metabolic activity in the muscle (Holmes et al. 1973). In fact, double‐muscled cattle have an increase in the proportion of fast twitch glycolytic fibres, resulting in a faster and more glycolytic phenotype (Girgenrath et al. 2005). Mutations in MSTN are responsible for DBM in other large animals including one case in humans. In the latter, Schuelke et al. (2004) observed that a G to A substitution (g.IVS1+5G>A) caused extraordinary muscling in a young boy, especially in the thighs and upper arms. No health problems were reported in the patient, and the testosterone and IGF‐1 levels were normal. In dogs known as ‘bully’ whippets, it was discovered that a 2‐bp deletion in the third exon of the MSTN is associated with the DBM phenotype. This deletion removes nucleotides 939 and 940 within exon 3 and leads to a premature stop codon at aa 313 instead of the normal cysteine, removing 63 amino acids from the predicted 375‐aa protein (Mosher et al. 2007). A gene‐targeting approach using the CRISPR/Cas9 system has been used to create MSTN‐null Beagles; although mutant dogs displayed the DBM phenotype, very little detail is available regarding their cellular phenotype (Zou et al. 2015). Due to the effects of MSTN on muscle mass, growth and other traits, the variations in MSTN expression levels in skeletal muscles are of great interest in the animal breeding field. Knowledge of null alleles and polymorphisms in the MSTN has been utilized to improve the selection of beef cattle and sheep (Georges 2010). The aim of this section of the review is to describe known DBM in livestock animals that harbour MSTN mutations. Mutations in the myostatin gene in cattle Monogenic determination of muscular hypertrophy in Belgian Blue cattle was first described in the 1980s (Hanset & Michaux 1985; Grobet et al. 1997). DBM was shown to be inherited as a single major autosomal locus that nevertheless was affected by several modifier loci manifesting in incomplete penetrance. The causal loss‐of‐function mutation in Belgian Blue MSTN, located on chromosome 2, was first reported by Grobet et al. (1997) followed shortly thereafter by McPherron & Lee (1997), who not only substantiated the findings of Grobet but also reported a missense mutation in exon 3 in the Piedmontese breed MSTN. Approximately 20 different types of genetic variants insertions and nucleotide substitutions, also known as single nucleotide polymorphisms (SNPs) have been identified in the bovine MSTN. Some of these genetic variants give rise to muscular hypertrophy by inactivation of the gene (Grobet et al. 1997). Variant alleles and inactive MSTN have a significant association with growth speed and favourite carcass traits, so these polymorphisms could be used in beef cattle to increase the quality and quantity of meat (Mirhoseini & Zare 2012). From the perspective of quality meat production, this is an outstanding trait, because these animals produce not only just more but also leaner and tenderer meat (Kobolák & Gócza 2002). The carcass and meat quality traits are superior in these animals because of a reduction in fat (decreased by 50%), an increase in muscle mass (by 20%), lower proportions of bone and also less connective tissue, which contributes to tenderness (McPherron & Lee 1997; Vincenti et al. 2007). However, dystocia‐related problems are often observed in double‐muscled cattle because hyperplasia occurs before birth, resulting in larger calves (Deveaux et al. 2001). Homozygous double‐muscled animals manifest more problems of dystocia than do heterozygotes. Therefore, to generate homozygous animals and at the same time keep costs down as well as reduce calf‐death probability, it is worth considering mating heterozygous animals (Bellinge et al. 2005). A summary of the detected genetic variants in cattle is reported in Table 1. Double‐muscled cattle breeds Belgian Blue The breed in which this muscular hypertrophy and its effects have been analysed most extensively is the Belgian Blue breed, which has been systematically selected for DBM to the point of fixation in many herds. Research by Grobet et al. (1997) revealed an 11‐bp deletion (g.821–831del11) in the open reading frame of the Belgian Blue MSTN allele that results in the loss of three amino acids (275, 276 and 277) and a frameshift after aa 274. The frameshift leads to a stop codon after aa 287. Work by Wegner et al. (2000) showed that semitendinosus from Belgian Blue was 1.6 times the weight of normal breeds due solely to an increase in muscle fibre number. Indeed, muscle fibre size from the Belgian Blue was actually smaller than in other breeds (Wegner et al. 2000). Furthermore, these animals have less collagen and connective tissue than do normal animals. The carcass fat content in these animals is significantly lower than in normal cattle, especially intramuscular fat (marbling), which is influenced by the DBM phenotype with a strong reduction of subcutaneous and internal fat tissues (Mirhoseini & Zare 2012). The results of many studies in fact have indicated that MSTN plays key roles in not only myogenesis but also adipogenesis. MSTN deletion and inhibition in animals lead mainly to increased muscle mass and reduced fat mass (Deng et al. 2017). In beef cattle production, crossing with Belgian Blue cattle shows that although the gene is recessive and monofactorial, its effect is apparent, even in heterozygous animals, due to its partial dominance (Kobolák & Gócza 2002). The same mutation has also been found in the Asturiana de los Valles (AV), a Spanish beef cattle breed. MSTN polymorphisms in the AV breed have been described, and its diffusion into the breed has been continuous for economic reasons (Grobet et al. 1997). Piedmontese In Piedmontese cattle, the DBM phenotype is an inherited condition associated with a g.938G>A substituion (in exon 3), which translates to p.Cys313Tyr in a highly conserved cysteine‐knot structural motif region of the protein. This is in the pre‐helix loop, a region known to be important for ALK4/5 receptor interaction (Cash et al. 2012). The mutation alters the function of MSTN, which disrupts a disulphide bridge that is essential for the correct conformation of the protein (Kambadur et al. 1997). This breed has been systematically selected for DBM to the point of fixation in many herds (>96% homozygous in the Piedmonte region in Italy), but variability in muscle mass is still present (Miretti et al. 2013). Several studies support the notion that the DBM phenotype, a partially recessive trait, causes relatively large effects on carcass conformation without a negative effect on calving, compared with animals with no copies of the mutated allele (Casas et al. 1998). Marchigiana The Marchigiana is one of the most important Italian beef cattle breed and renowned for its large body size, high weight daily gains and superior carcass dressing percentage. Marchigiana breed individuals have a g.874G>T mutation in exon 3, translating to p.Glu291Xaa in MSTN. This point mutation has a remarkable effect on the MSTN gene, as it changes a codon for glutamic acid into a stop codon (Marchitelli et al. 2003). In Marchigiana, as in the other DBM breeds, the MSTN genotypes yield three different and distinct phenotypes. Homozygous G/G individuals display the normal phenotype, whereas those with the T/T genotype manifest as a double‐muscled body shape while maintaining their small frame and are frequently associated with skeletal defects and serious survival problems due to macroglossia and hypoplasia of the heart, lungs and other vital organs. Individuals with a heterozygous genotype (G/T) display a well‐muscled and large body structure and excellent conformation without any of the above‐mentioned defects. Therefore, the heterozygous animals are frequently selected as sires (Cappuccio et al. 1998). Moreover, heterozygous animals show better meat quality than do animals with a normal genotype (Vincenti et al. 2007). Therefore, they could be useful for breeders in planning matings to obtain a higher number of heterozygous animals. Obviously, this is possible only if the genotype at the MSTN locus of each animal is available. Additionally, two different SNPs have been found in the promoter region—g.–371T>A and g.–805G>C (genomic numbering relative to the ATG start codon)—although Sarti et al. (2014) reported that these substitutions may not be useful for consideration in selection criteria, because there is no correlation with productive traits or due to their homozygous genotype. Other cattle breeds An 11‐bp deletion (c.821_831del11), resulting in a truncation of the bioactive C‐terminal domain of the protein has been found in Blonde d'Aquitaine, Limousine Parthenaise and Rubia Gallega breeds of cattle (Kambadur et al. 1997; Dunner et al. 2003). A recent study identified an unexpected mutation in the MSTN gene in Blonde D'Aquitaine cattle (Bouyer et al. 2014). The mutant allele is highly expressed, leading to an abnormal transcript consisting of a 41‐bp inclusion between the exons 2 and 3 with a premature termination codon predicted to translate into a protein lacking the entire bioactive region. An additional transversion mutation (g.433C>A) in the Limousine breed that was shown to be functionally associated with the increased muscle mass and carcass yield without any associated reproductive disadvantages has been described (Sellick et al. 2007; Esmailizadeh et al. 2008; Vankan et al. 2010). As in Piedmontese cattle, a g.938G>A transition has been reported in Gasconne (Kambadur et al. 1997; Dunner et al. 2003). An indel (c.419_421del7ins10) was reported in Maine‐Anjou cattle, resulting in a premature stop codon in the N‐terminal latency‐associated peptide at aa 140 (McPherron & Lee 1997). Additionally, a c.676G>T substitution, also causing a premature stop codon in the same N‐terminal latency‐associated peptide (p.Glu226Thr), was identified in the same breed (Grobet et al. 1997). Charolaise and Limousine have a c.610C>T transition, yielding a premature stop codon in the N‐terminal latency‐associated peptide (p.Gln204Xaa) (Cappuccio et al. 1998). In addition to the genetic variants found in Bos taurus, 14 polymorphisms (three in exon 1, seven in exon 2 and four in exon 3) have been reported in the coding part of the MSTN in the Nellore cattle (Bos indicus) genome. However, whether or not these polymorphisms are functional mutations still remains to be elucidated (Grisolia et al. 2009). Double muscling in sheep The MSTN gene is located at the end of the long arm (2q32.2 locus) on chromosome 2 in sheep (Ovis aries) (Bellinge et al. 2005). During the past decade, a total of 77 MSTN SNPs have been reported in various sheep breeds such as Texel, Norwegian Spælsau, commercial New Zealand (NZ) sheep breeds and Latvian Darkhead (Kijas et al. 2007; Sjakste et al. 2011; Han et al. 2013), and the majority of these SNPs are located in the non‐coding regions of the gene. The exceptions are a 1‐bp deletion (MSTN:c.960delG) in Norwegian White Sheep and 1‐bp insertion (c.120insA) in NZ Romney (Boman et al. 2009). Lastly in 2018, Trukhachev et al. described for the first time eight variations in non‐coding regions of MSTN in the Stavropol Merino, a breed used for meat production in Russia. A summary of the detected genetic variants in sheep is reported in Table 2. Texel sheep Belgian Texel sheep muscle fibres show enlargement and therefore can be considered to have fibre hypertrophy. Texels are utilized extensively as a terminal crossbreed because of their exceptional conformation and potential to produce higher yielding carcasses with increased lean and decreased fat content (Leymaster & Jenkins 1993). Analysing the MSTN gene revealed no nucleotide differences in the coding regions between double‐muscled and normally muscled animals (Kijas et al. 2007). This suggests that genetic variation located outside the coding regions plays a more important role in the regulation of muscle development, in contrast to cattle, for which MSTN loss‐of‐function variants have been found within the three coding exons (Grobet et al. 1997). Quantitative trait locus analysis in Texel sheep characterized a variant (g.6723G>A) in the 3′‐UTR (untranslated region) of MSTN on chromosome 2 that has an effect on muscle mass. This creates a target site for miR1 and miR206, microRNAs that are highly expressed in skeletal muscle (Kijas et al. 2007). Other genetic variants have also been found including c.*1232A, g.391G>T and 18 other SNPs: g.2449C>G, g.2379C>T, g.1405A>T, g.1402G>A, g.1214C>T, g.1129C>T, g.41A>C, g.39T>C, g.474C>T, g.613T>C, g.616G>A, g.619T>C, g.622T>C, g.632G>T, g.696C>T, g.3135C>T, g.4036A>C and g.4044C>T (Kijas et al. 2007). Norwegian sheep The DBM phenotype in Norwegian White sheep has been described as extraordinary over‐development of muscles, particularly on the hindquarters. Investigations have shown that these animals have not only extremely low levels of subcutaneous fat but also decreased internal fatty tissues. The DBM animals have lower bone mass compared with wild type animals. Sequence analysis revealed a 1‐bp deletion in MSTN (c.960delG) in DBM individuals. The deletion of this G residue disrupts the reading frame from aa 320 onward and produces a premature stop codon at aa 359 (compared to aa 375 in wild type animals) (Boman & Våge 2009). The same MSTN 3′‐UTR mutation (c.*2360G>A) identified in Texel sheep was also found in the Norwegian breed but with a less profound effect (Boman & Våge 2009). However, a similar phenotype of increased muscle mass and fat was found in Norwegian Spælsau sheep. The sequencing of the MSTN coding region revealed a c.120insA insertion in DBM animals. The insertion of an adenine residue disrupts the reading frame from aa position 40 onward and generates a premature stop codon at aa position 49 (Boman & Våge 2009). New Zealand sheep A comprehensive investigation of polymorphisms in MSTN in a diverse range of NZ sheep breeds (Romney, Coopworth, Corriedale, Dorper, Perendale, Suffolk, Merino, Dorset Down, Poll Dorset, Texel and other NZ cross‐bred sheep) was performed using polymerase chain reaction‐single strand conformational polymorphism (PCR‐SSCP) analysis and DNA sequencing. A total of 28 nucleotide substitutions were identified from nucleotide c.–1199 (in the promoter region) to c.*1813 in the 3′‐UTR. Of these, three were located in the promoter region, three in the 5′‐UTR, 11 in intron 1, five in intron 2 and five in the 3′‐UTR. Ten new substitutions have been reported: c.–959C>T, c.–784A>G, c.373+563A>G, c.373+607A>G, c.374–654G>A, c.374–54T>C, c.748–54T>C, c.*83A>G, c.*455A>G and c.*709C>A (Han et al. 2013). The other 18 substitutions had been reported previously. These include c.101G>A, which had already been found in NZ Romney by Zhou et al. (2008) and also in Merino, Corriedale and NZ cross‐bred sheep (Clop et al. 2006; Kijas et al. 2007). In NZ Romney, two additional SNPs—c.–2449G/C and c.–2379T/C—were detected (Wang et al. 2016). The c.*123A variant observed in NZ cross‐bred sheep was also reported in Texel (Kijas et al. 2007); Charollais sheep from Britain (Hadjipavlou et al. 2008); and White Suffolk, Poll Dorset and Lincoln breeds from Australia, and this variant showed significant association with the DBM phenotype as well as with the other substitution c.373+18T>G reported in Texel sheep (Clop et al. 2006). Other sheep breeds Zel sheep'','' a meat breed in northern Iran, has a polymorphism in intron 2 as does the Iranian Baluchi sheep (Dehnavi et al. 2012). Three polymorphic sites in Indian sheep have been identified in the 5′‐UTR, exon 1 and exon 2 regions. Both SNPs in the exonic region were found to be non‐synonymous. The genetic variants c.539T>G and c.821T>A were in exons 1 and 2 respectively (Pothuraju et al. 2015). None of these genetic variants is significantly associated with the DBM phenotype. Myostatin polymorphisms in goat Allelic variation in the goat MSTN has been investigated in several studies. A 5‐bp indel (c.12561260delinsTTTTA) was identified in the 5′‐UTR region in Boer, Matou, Haimen and Nubi goat breeds, and a substitution (g.1388T>A) in exon 1 was detected only in Boer (Zhang et al. 2012a,b). Two novel SNPs were also identified in Boer and Anhui white goat: g.197G>A, a substitution located in the 5′‐UTR, and g.345A>T in exon 1 (Zhang et al. 2013). A thorough investigation was conducted in 22 different goat breeds (Inner Mongolia Cashmere, Liaoning Cashmere, Taihang Mountain, Chengde Polled, Jining Grey, Tibetan, Chengdu Brown, Jianchang Black, Guizhou White, Guizhou Black, Longlin, Duan goat, Leizhou, Matou, Yichang White, Shannan White, Nanjiang Brown, Angora, Toggenburg, Nubian, Saanen and Boer goat), and a total of eight SNPs were detected (g.1980A>G, g.1981G>C, g.1982A>G, g.1984G>T, g.2121A>G, g.2124T>C, g.2174G>A and g.2246A>G) (Li et al. 2006). Recently, Nguluma et al. (2018) detected a polymorphic site (g.298T>C) in the Boer goat population; the authors concluded that the potential association of this polymorphism in MSTN with growth performance could not be confirmed and that other genes for growth could be responsible for the observed variation. A summary of the detected genetic variants in goat is reported in Table 3. Myostatin polymorphisms in horse Hosoyama et al. (2002) isolated and sequenced MSTN cDNA from a Thoroughbred horse that was mapped to chromosome 18. Mutations in the equine MSTN have been identified and are associated with racing phenotypes influencing racing performance and muscle fibre proportions (Petersen et al. 2013). Dall'Olio et al. (2010) sequenced 16 horse breeds (Rapid Heavy Draft, Noric, Bardigiano, Haflinger, Lipizzan, Murgese, Tolfetano, Uruguayan Creole, Italian Saddle, Maremmano, Quarter Horse, Salernitano, Andalusian, Ventasso, Italian trotter, Thoroughbred horse), revealing seven SNPs: two transitions that were located in the promoter region 646 (GQ183900:g.26T>C) and 156 (GQ183900:g.156T>C) bp upstream from the start codon and are associated with breeds of different morphological types. The g.26T>C SNP was polymorphic in six of 16 breeds with higher observed frequency of the g.26C allele. The g.156T>C polymorphism was detected in 11 of 16 breeds and was identified in a homozygous condition in a few Bardigiano, Haflinger, Noric, Rapid Heavy Draft and Uruguayan Creole horses (Dall'Olio et al. 2010). The other five SNPs were in intronic regions, four localized in intron 1 and one in intron 2. Three of the SNPs in intron 1 (g.1634T>G, g.2115A>G and g.2327A>C) were also identified in Thoroughbreds (Petersen et al. 2013). One polymorphism (g.2115A>G) has been associated with sprinting ability and racing stamina in Thoroughbred horses. The association between MSTN and horse racing performances was further evidenced by Binns et al. (2010) and Tozaki et al. (2010). Subsequently, 15 Chinese breeds were studied to select the best Chinese domestic breed for evaluating potential racing performances (Li et al. 2014). These studies found six different SNPs in MSTN: two SNPs (g.26T>C and g.156T>C) in the promoter region, two (g.587A>G and g.598C>T) in the 5′‐UTR region and two (g.1485C>T and g.2115A> G) in intron 1 of the equine MSTN. SNPs g.587A>G and g.598C>T were novel, whereas the others had been previously reported (Petersen et al. 2013). Baron et al. (2012) described a genetic variant in exon 2 in some horse breeds. In fact, they identified a g.2279A>C substitution in Arabians horses and a g.2478G>C substitution in the Soraia breed horse. Five polymorphisms (g.66495826T>C, g.66495696T>C, g.66493737T>C, g.66495254C>T and g.66490010T>C) were recently observed in four Polish breeds (Arabians, Polish Konik, Hucul and Polish Heavy Draft) (Stefaniuk et al. 2016). The g.66495254C>T polymorphism (also referred to as g.598C>T; Li et al. 2014) has been described in Chinese horse breeds as well as in Polish Konik and Arabian horse breeds. The g.66493737C>T polymorphism, known to predict optimum distance in Thoroughbred horses, has been identified in four breeds in Egyptian bloodlines (Bower et al. 2012), which were introduced to Polish bloodstock through Egyptian stallions. The g.66495326_66495327ins227 insertion has been described for the first time in MSTN in Thoroughbred horses. Recently, it was found in the American Quarter Horse (Petersen et al. 2013) and in the Uruguayan Creole breeds (Dall'Olio et al. 2014). In the Quarter Horse breed, this MSTN insertion is connected with changes in gluteus medius muscle fibre proportions. The higher myosin heavy chain 2B fibre type (fast contracting) is in line with pressure selection in the Quarter Horse breed for racing performance (Petersen et al. 2013). A summary of the detected genetic variants in horse is reported in Table 4. Myostatin polymorphisms in pig Jiang et al. (2002) reported three SNPs in the porcine MSTN gene: T>A, G>A and C>T (accession nos. AF393396, AF393397 and AF393398 respectively), in the promoter, intron 1 and exon 3 respectively. Only one mutation (MSTN:g.383T>A) was associated with average daily gain during the growing period (from 60 to 100 kg of live weight) in Yorkshire pigs. Furthermore, bodyweight in pigs with the heterozygous mutation (no aa was found) was greater (Jiang et al. 2002). Stinckens et al. (2008) compared the MSTN sequence of Belgian Piétrain, which shows a heavily muscled phenotype, with five other breeds (Piétrain, Landrace, Large White, Meishan and Wild Boar). Fifteen polymorphic loci were found, three of which were located in the promoter region (g.435G>A, g.447A>G and g.879T>A), five in intron 1 and seven in intron 2. The g.879T>A polymorphism appears only in Chinese Meishan pigs, whereas the g.447A>G polymorphism located in the porcine MSTN promoter had a very high allele frequency in the Piétrain pig breed. The g.447A>G variant, which is associated with the expression of the porcine MSTN, occurs at the putative myocyte enhancer factor 3 (MEF3) binding site on the negative DNA strand, and the mutation disrupts a putative MEF3 binding site (Stinckens et al. 2008). However, these results suggest that naturally occurring MSTN genetic variants identified thus far in pigs do not have significant association with muscle phenotypes. Nevertheless, recent work using an experimental approach has shown the role of MSTN in the development of muscle in pigs. Qian et al. (2015) generated MSTN‐deficient Meishan pigs using zinc finger nuclease technology coupled with somatic cell nucleus transfer. The resulting offspring show a remarkable DBM phenotype, especially pronounced in the hindquarters. Muscle in the MSTN‐null pig increased mass by 50–100%. Incredibly, the muscle fibre size in the null pigs was smaller than in the wild type. All the increase in mass could be attributed to fibre hyperplasia whereby some muscles from the null animals had twice the number of fibres compared to the wild type. The animals displayed good overall health. As the technology employed did not involve the introduction of any genetic material into the genome (e.g. selection markers), Qian et al. (2015) suggest that it is essentially the same as double‐muscled cattle that are used for human consumption. A summary of the detected genetic variants in pigs is reported in Table 5. Myostatin polymorphisms in rabbit Fontanesi et al. (2011) investigated the variability of the effects of MSTN polymorphisms on rabbit production traits. Four single SNPs were identified by comparative sequencing of 14 rabbits representing breeds or lines having different conformation and muscle mass: one rare synonymous SNP in exon 1 (c.108C>T), one synonymous SNP in exon 2 (c.713T>A), one SNP in the 3′‐UTR (c.*194A>G) and another SNP in intron 2 (c.747+34C>T) in Belgian Hare, Burgundy Fawn, Checkered Giant and Giant Grey. In commercial hybrids, Qiao et al. (2014) detected a SNP (g.476T>C) in the 5′ regulatory region, but no mutation sites were detected in the exons. A correlation analysis showed that the SNP was associated with increased liver and carcass weight. These results suggest that SNPs in the upstream regulatory region of the MSTN are beneficial to rabbit soma development and that the variants can be used as molecular markers for the selection of meat quality in rabbits. Sternstein et al. (2014) found polymorphisms in the MSTN in the Giant Grey and NZ White breeds. Comparative sequencing of these breeds revealed two SNPs located in the regulatory region (c.–125T>C) and in intron 1 (c.373+234T>C) of the rabbit MSTN. A summary of the detected genetic variants in rabbit is reported in Table 6. Myostatin polymorphisms in poultry In chickens, the MSTN gene maps to 7p11 (Sazanov et al. 1999) and, similar to mammals, is composed of three exons (373, 374 and 1567 bp respectively) and two introns. Gu et al. (2003) showed that poultry MSTN not only regulates skeletal muscle development but also participates in fat metabolism and disposition. They identified seven SNPs: five in the 5′‐regulatory region (g.167G>A, g.177T>C, g.304G>A, g.322A>G and g.334C>T) and two in the 3′‐regulatory region (g.7263A>T and g.6935A>G) of different chicken lines. Ye et al. (2007) studied the association of MSTN polymorphisms with mortality rate, growth, feed conversion efficiency, ultrasound breast depth, breast percentage, eviscerated carcass weight, leg defects, blood oxygen level and hen antibody titre to the infectious bursal disease virus in three commercial broiler chicken lines. MSTN had pleiotropic effects on broiler performance. This conclusion was reached by the discovery of 14 SNPs: seven genetic variants in exon 1 (g.2100G>A, g.2109G>A, g.2244G>C, g.2283A>G, g.2346C>T, g.2373C>T, g.2416A>G), one in exon 2 (g.4842T>G), three in exon 3 (g.7434C>G, g.7435A>G, g.7436C>A) and three in introns 1 and 2 (g.4405G>C, g.4405A>T and g.4954A>G). As the main function of MSTN is the regulation of skeletal muscle growth, Ye et al. (2007) deemed that the non‐synonymous SNP g.4842T>G is associated with an amino acid change in MSTN and could be responsible for variability in body weight. The Bian chicken breed, raised for dual purposes, is an important Chinese breed and has a g.234G>A SNP in exon 1 of the MSTN (Zhang et al. 2012a,b). Other Chinese chicken breeds (Jinghai, Youxi and Arbor Acre) have been shown to have four new variants (g.326A>G, g.334C>G, g.1346C>T and g.1375G>A), located in the 5′‐regulatory region (Zhang et al. 2012a,b). Further studies on growth traits have shown that the SNPs in chicken MSTN may affect abdominal fat weight and percentage, breast muscle weigh and percentage, birth weight and adult weight (Zhang et al. 2012a,b). Zhiliang et al. (2004) identified three SNPs in the 5′‐regulatory region and two SNPs in the 3′‐regulatory region and that these differed in allele frequencies between breeds. They found that in an F2 generation from a cross of broiler and silky chickens, homozygous genotypes AA and BB at a locus in the 5′‐regulatory region had higher abdominal fat weight and abdominal fat percentage than did those with the AB genotype (Zhiliang et al. 2004). The upstream promoter region of MSTN was analysed in Wenshang Luhua chicken DNA. Thirteen E‐boxes were identified upstream of MSTN, and the E‐box polymorphisms were explored for the first time (Hu et al. 2013). Other interesting studies were carried out on ducks to investigate the association of polymorphisms in MSTN with slaughter traits, breast muscle weight, breast muscle percentage, leg muscle weight and leg muscle percentage. Analysis of the 5′‐regulatory region of MSTN showed that polymorphisms (g.753G>A, g.658G>T and g.235G>C) were associated with breast muscle percentage and abdominal fat rate (Lu et al. 2011). Furthermore, Xu et al. (2013) studied polymorphisms in Pekin duck and identified three significant variations. The first is a c.129T>C subsitution located in the open reading frame and revealed an association with breast muscle thickness. The second SNP (c.708T>C) was also located in the open reading frame, and the third (c.952TA substitution in exon 3 of MSTN is correlated with abdominal fat rate (Liu et al. 2012). In Sansui duck, six SNPs were identified in the first and the third exons (g.106G>A, g.120A>G, g.159G>A, g.5368G>A, g.5389A>C and g.5410G>A) with four loci seemingly associated with leg muscle weight, leg muscle percentage and dressing percentage (Zhao et al. 2016). A summary of the detected genetic variants in poultry is reported in Table 7. Myostatin and future implications According to some investigators, MSTN variants are the main cause of hypertrophy, with lesser roles played by other gene variants (Kobolák & Gócza 2002). Inactivation of MSTN has therefore been proposed as a strategy for improving muscle growth of food animals and treating human diseases associated with muscle weakness and dystrophy (Chen & Lee 2016). Research, especially on mice, has highlighted the potential of manipulating MSTN signalling to promote muscle growth. In null mutants of this species, some muscles are approximately three times their normal weight. Impressive as they are, muscle enlargement in large mammals carrying a null mutation in the same gene, to our knowledge, does not approach this level of muscle growth. Therefore, it is important to ascertain the molecular basis underpinning these different responses with a view of translating these findings into increased meat production. One picture that emerges through this review is that mutations that compromise MSTN function have a consequence during development and give rise to supernumerary muscle fibres (hyperplasia). However, one of the clear differences between mice and large animals (cattle and pigs) is the post‐natal phenotype. Mice show considerable fibre hypertrophy, whereas both cattle and pigs display no increase in fibre size. These findings need to be used as a benchmark for future work on DBM in large animals. First and foremost is the need to understand the basis of muscle growth in large mammals. It is very important to use the correct terms to describe the phenotype of animals, as often this can lead to misinterpretations regarding mechanism. Often DBM animals are referred to as being ‘hypertrophic’; however this could infer fibre enlargement. As we have discussed, especially in the case of cattle and pig, there is no fibre enlargement. We suggest that accurate mechanistic descriptors be used when they have been precisely established, and without this proof, a more generic term needs to be applied. We suggest the use of the four following terms: (i) muscle enlargement through hyperplasia, (ii) muscle enlargement through hypertrophy, (iii) muscle enlargement through hyperplasia and hypertrophy and (iv) muscle enlargement through unknown cellular mechanisms. Research is required to understand the mechanisms that underpin the role of MSTN in post‐natal muscle development in mammals—to answer the question why, in the absence of MSTN, fibres from mice undergo enlargement whereas those from large mammals do not. For a number of years, the naturally occurring variants in cattle were the only reference model for large animals lacking MSTN. The lack of fibre hypertrophy was usually explained by the presence of a secondary (to date unidentified) modifying mutation that interfered with the post‐natal effect but spared the pre‐natal phenotype. However the work by Qian et al. (2015) in the pig, which targeted only MSTN, undermines the modifying gene idea. Therefore, loss‐of‐function mutations in both small and large animals lead to hyperplasia. However, it is only in mice that the mutation has an effect on muscle fibre size, where it presents as hypertrophy. Clues to resolving this issue come from recent work in monkeys, which has shown that MSTN and activin act synergistically to inhibit fibre hypertrophy during adult life (Latres et al. 2017). Based on these findings, we suggest that muscle fibres of both cows and pigs are sensitive to myostatin/activin signalling in a manner similar to monkeys. However, the issue that still needs to be resolved is why fibres in adult cows and pigs fail to enlarge in the absence of MSTN. The most parsimonious explanation is that there is a partial redundancy relationship between MSTN and activin; in the absence of MSTN, the expression levels of activin become elevated to such a degree that in cows and pigs the latter can completely cover the loss of the former. Examples of gene expression compensation by related molecules, similar to our proposal, abound in mammalian biology (Barbaric et al. 2007). One of the best examples comes through the investigations of MRFs, whereby genetic inactivation of MyoD results in an up‐regulation of the Myf5 (Rudnicki et al. 1992). The hypothesis outlined above has a number of important implications. Our assertion of why the relationship between MSTN and activin in cows and pigs is only partial and not complete comes from the fact that loss of MSTN has some phenotypic consequence (hyperplasia). Therefore, compensation through an up‐regulation of activin expression cannot have occurred during pre‐natal life. The second implication is that if there is a redundancy mechanism in mice, it must be very muted because these animals develop a profound phenotype both during pre‐natal and adult life. Our suggestions can be validated by quantifying the levels of MSTN and activin at different developmental stages in both large and small animals, an avenue now possible following the development of specific ELISAs for MSTN and activin (Latres et al. 2017). For the meat industry and for the human health sector, for which the focus is on muscle growth, the hypothesis outlined here advocates a strategy of dual MSTN and activin antagonism to promote the growth of the tissue. This could be achieved through the use of a combination of molecules that specifically antagonise the activity of MSTN and activin (antibodies or protein‐specific propeptides) or a single protein that acts at a signalling convergence point at the receptor level through the deployment of a ligand trap or blocking antibody (Lach‐Trifilieff et al. 2014; Omairi et al. 2016). Moreover, it will be very interesting to better understand the role of MSTN in adipogenesis for beef production; in fact, Deng et al. (2017) reported that muscle and adipose tissue develop from the same mesenchymal stem cells, and researchers have found that MSTN is expressed in fat tissues and plays a key role in adipogenesis. Finally MSTN is a prime target for transgenic approaches aimed at enhancing meat production in livestock (Georges 2010). Possible strategies for this outcome include the generation of MSTN knock‐out animals. Also, more elaborate transgenic approaches, such as targeting post‐natal or sex‐specific inhibition of MSTN, need to be considered. Wang et al. (2017) reported the successful application of the CRISPR/Cas9 system to engineer the goat genome through micro‐injection of Cas9 mRNA and sgRNAs targeting MSTN in goat embryos. They demonstrated the utility of this approach by disrupting MSTN, resulting in enhanced body weight and larger muscle fiber size in Cas9‐mediated gene modified goats. MSTN activity can also be modified using non‐genetic approaches, for example using blocking antibodies or ligand traps. Conclusions One picture that emerges through this review is that mutations that compromise MSTN function have a consequence during development and give rise to supernumerary muscle fibres (hyperplasia). However, one of the clear differences between mice and large animals (cattle and pigs) is the post‐natal phenotype. First and foremost there is the need to understand the basis of muscle growth in large mammals. This review provides the landscape of the genetics of DBM in mammalian species and chicken and demonstrates the huge number of genetic variants present in animals of commercial interest. It also highlights areas where greater research is required for progress to be made concerning the role of MSTN in the regulation of muscle development in economically important animals. Knowledge of null alleles and polymorphisms in MSTN are of great interest in the animal breeding field and could be utilized to improve the selection for meat production in livestock animals. Acknowledgements The authors want to thank the three anonymous referees for their valuable comments and constructive suggestions. Conflict of interest The authors have no conflict of interest to declare.